Research Load Identify Algorithm Of CPT System Base On Neural Networks

Contactless Power Transfer (CPT) technology bases on magnetic coupling,and realizes by applying modern control theory deliver power from stationary source to movable loads. In order to minimize the volume and weight of the system. This system employs high-frequency converting technology and supplies power to loads. It's a new way of power transmission which is safe, reliable and overcomes disadvantages of traditional power transfer methods.When CPT system is on the normal operation,the variation of load will cause the system operating frequency to occur drifts,and deviate resonance frequency scope. And thus it will cause the system robust variation,and affect the system's normal operation. It will even cause the system to collapse in the serious case. After making the CPT system load recognition algorithm's research,people realize that promptly adjusting the control quantity according to the recognition result,enables the system operating frequency to return to the resonance frequency scope , thus enhances system's robustness,and has the system on the normal operation again.This paper has proposed and studied a study of load identify algorithm of contactless power transmission that bases on the neural network,through the load identifying,to improve the related issues caused by load changing.This paper introduces contactless power transmission system on principle,the functions of various components,analyzes the frequency shift and output power reduce caused by load changing. For different nature of loads,different mathematical models are adopted for analysis. The nature of load corresponding to the original trace slip-output and the relationship between the frequency; Introduction the fast Fourier and neural network theory and knowledge。Comparing the results of the two different theory by using Matlab and Pspice,the result shows that load identifying algorithm based on neural networks more precisely than that based on the fast Fourier transform,and it has achieved the expected results.